Dimensional Check of Catalyst Tubes

to Assess Remaining Life

In this article, a procedure f or measuring the remaining life of catalyst tubes based on diameter
measurements, specifically ID bulging, is outlined. Our experience with this method when applied
to catalyst tubes that have been in use for more than 10 y ears is discussed, with examples drawn
from nine different plants.

T. Shibasaki, T. Mohri, and K. Takemura

Chiyoda Corporation, Tsurumiku Yokohama 230-8601, Japan

length to find the catalyst tube's temperature profile.

C
atalyst-tube life is consumed by creep damage,
which depends on stress, time, and tempera- « Measurement of OD and tube wall thickness at the
ture. The internal pressure on catalyst tubes same position to evaluate the amount of creep damage.
and the service time are equivalent within the same • Observation of microstructures from the outer sur-
plant. However, each tube has a slightly different max- face side of the catalyst tube by the replica method to
imum temperature, which determines the remaining confirm the metal temperature and the creep damage.
life of catalyst tube. The amount of creep damage is
evaluated by dimensional examination. Measurement of Hardness Distribution
Figure 1 shows the change of accumulated number
of catalyst tubes for each tube material. In this time, Catalyst tubes are made of heat-resistant cast alloy.
the use period of IN519 (24Cr-24Ni-Nb-0.3C iron The hardness of heat-resistant cast alloy changes along
base cast alloy) catalyst tubes is more than 12 years, with temperature and time. In the case of used catalyst
which corresponds to about 100,000 h. We presented tubes within the same furnace, the service time is the
the preliminary assessment on the remaining life of same. The hardness change in the catalyst tubes along
EST519 catalyst tubes at the Vancouver ammonia sym- the tube length shows the temperature profile during
posium in 1994 (Shibasaki et al., 1995). After that, the operation.
many IN519 catalyst tubes that were 12 years old were Figures 3 and 4 show the results from measurement
investigated to estimate the remaining tube life (Figure of Vickers hardness distribution for IN519 catalyst
2). Table 1 shows the design conditions of the investi- tubes that were used hi a hydrogen plant for 20 years
gated IN519 catalyst tubes. and in an ammonia plant for 19 years, respectively.
We propose a procedure to assess the remaining life The horizontal axis indicates the location of the hard-
of catalyst tubes by dimensional check. This procedure ness measurement (0% shows the top of catalyst tube,
consists of the following three steps. where it is outside the firebox, and 100% shows the
• Measurement of hardness distribution along the tube bottom of catalyst tube).

AMMONIA TECHNICAL MANUAL 156 1999

 HK40 IN519 tff-Nb.TI

; IN5I9 HP-Nb.T! • : Catalyst tube »as sampled to investigate.

Figure 1. Change in material for catalyst tube Figure 2. Used portions of investigated IN519 catalyst
(from 1965 to 1999). tubes.

—O—Canter -Oulsidoj

Hardness Hardness
(Hv) (Hv)

20 40 60 80 100 20 40 €0
Location (X. 0:Top — 100:Botto») Location <X. 0:Top — IMrBottw)

Figure 3. Hardness distribution along tube length Figure 4. Hardness distribution of used IN519 cata-
of used IN519 catalyst tube for 20 years in plant A. lyst tube for 19 years in plant B.

Table 1. Design Conditions of the ÏN519 Catalyst Tube Investigation

Plant
A B C E F H J L M
Used Period (yr) 20 19 15 15 15 14 14 12 14
Design Temperature (K) 1,166 1,153 1,182 1,168 1,188 1,173 1,185 1,191 1,183
Design Pressure (MPa) 3.2 4.1 2.9 2.8 2.8 2.7 2.7 2.9 2.7
Hydrogen (H) or Ammonia (A) H A H H H H H H H

AMMONIA TECHNICAL MANUAL 157 1999

 [ —o— (0°- —O— ( W - i
154

153.5

Outsida
16.5
Diawtsr Tube Mal I
On) 153 Thickness
ta)
16
152.5

152 15.5
20 40 60 80 100 •20 40 60 80 100
Location OS. 0:Tcp — 100:Bott») Location (X. OITop — 100:Botton)

Figure 5. OD distribution of used IN519 catalyst Figure 6. Tube-wall thickness distribution of used
tube for 19 years in plant B. IN519 catalyst tube for 19 years in plant B.

| —0- (O«- 18« Off- 270)

121

120.8 /\
153.5

.InsldB
,-L. Outside
Disaster
ianeter (Ml) 153
120.4

152.5
120.2

120
20 40 60 80 100 20 40 60
Location OS, 0:Top — 100:Bottoo) Lout ion (X, 0:Top — 100:Bot to»)

Figure 7. ID distribution of used IN519 catalyst Figure 8. OD distribution of used IN519 catalyst
tube for 19 years in plant B. tube for 20 years in plant A..

The temperature at the 0% portion is lower than 550°C distribution density (the number of carbides in the unit
and the hardness of its portion shows as cast condition. area) decreases and the shape of carbides becomes
On the other hand, in the cell corresponding to 4% massive. This mitigates the level of hardness.
through 100% portions, the temperature of catalyst Therefore, the distribution of hardness corresponds to
tube gradually increases in the downwards direction. the temperature profile of catalyst tube. The softest
Hardness changes along with the morphology of the portion corresponds to the highest tube skin tempera-
precipitated carbides due to aging. At around the 10% ture portion. It is important to find the highest temper-
portion, many fine carbides precipitate in the matrix of ature portion, since the creep damage in this portion is
the catalyst tube, and the hardness indicates the maxi- most severe and remaining life is shortest.
mum level. The precipitated carbides coalesce and The shape of hardness distribution is almost identi-
coarsen as they progress further down the tube. The cal in plants A and B, but the hardness levels in the

AMMONIA TECHNICAL MANUAL 158 1999

 faces. In fact, the original machining scratches are still
present after long periods of use in many catalyst
tubes. The increase in ID correlates to the creep
bulging.
All of the catalyst tubes in the firebox are measured
at l m intervals from the top to the bottom of the tube
on the burner side (0-180) and the row side (90-270),
that is, the side of the tube facing the next tube in the
row.
The ID is calculated by subtracting both sides of
tube wall thickness from the OD. Figures 5, 6, and 7
show the results of measurements of the OD, tube wall
20 40 60 thickness, and the calculated IDs along the tube length
Location (X, 0:Top — 100:Bott<n)
for IN519 catalyst tubes that were used in an ammonia
plant for 19 years. It is difficult to measure the pattern
Figure 9. lube-wall thickness distribution of used
of dimension change caused by creep from the result
IN519 catalyst tube for 20 years in plant A.
of the OD measurement (Figure 5). On the other hand,
softest portions are different. The difference in the it is easy to evaluate the change of dimension from ID
hardness level is caused by the service temperature of measurement (Figure 7).
the catalyst tubes. The metal temperature of the cata- In the actual catalyst tubes, there are some manufac-
lyst tube in plant A is higher than that in plant B. The turing deviations in size for each unit spun-cast tube.
difference of service temperature influences the The manufacturing deviation of the OD is larger than
behavior of creep deformation of catalyst tubes. that of the ID because the outside surface of a unit
It is best to measure the hardness distribution for all spun-cast tube is as-cast condition and the inside sur-
of the catalyst tubes. If this is not practical, a mini- face is machined. Figures 8, 9, and 10 show the results
mum of two kinds of tubes should be measured. One of OD measurements, tube-wall thickness, and ED cal-
tube should be selected randomly from the tubes used culations for the IN519 catalyst tube after 20 years of
at the average service condition. The other should be service in plant A. The butt weld joint is located at the
selected from the tubes that have undergone some- location 65% down the tube. A big difference in the
thing abnormal, such as severe bending or overheat- OD was observed at the butt weld (Figure 8). The
ing. manufacturing deviation was compensated for by mea-
suring the dimensions on both sides of the weld joint
Measurement of OD and Tube Wall and comparing the difference of the IDs. If the differ-
Thickness ence between the IDs was 0.2 mm or more, the down-
stream-tube ID was corrected and then added or sub-
EST519 catalyst tubes bulge due to creep phenomena tracted to the difference. A reasonable continuity in the
after long operation periods, and both the OD and the ID IDs between two unit tubes is obtained as shown in
increase. The change of dimension corresponds to degree Figure 10. This is another reason to measure the inside
of creep damage, and the remaining tube life can be esti- rather than the outside diameter.
mated by measuring this dimensional change. The change of dimensions was calculated on the basis
The outside surfaces of catalyst tubes are exposed to of the initial dimensions of the tube, rather than the
the flue gas and attacked by oxidation, a process that dimensions of drawing of the tube. Usually, the portion
ultimately reduces the OD. Therefore, the OD does not outside of the firebox retains its original dimensions.
reflect bulging from creep damage. The inside sur- Figures 11 and 12 show the change of IDs along the tube
faces are exposed to the reformed gas and are hardly length for the used catalyst tubes in plant A and plant B.
oxidized or corroded compared with the outside sur- The change of ID increases as it goes downstream.

AMMONIA TECHNICAL MANUAL 159 1999

 <0*- 130»
(o"- - (90»- 270)
I.——fill

1.5
Change of
Inside
Insida Diameter
Diaœter 00 l
(n») 128

20 40 80 100 20 40 60
Location (S, 0:Top — 100:Bottom) Location (X, OCTop — 100:Bottco)

Figure 10. ID distribution of used IN519 catalyst Figure 11. Change of ID distribution of used IN519
tube for 20 years in plant A. catalyst tube for 20 years in plant A.

|—o—Oam»ol irrar d i are la r (0*- 180? | -Hardness (Center) j

| -o— (0*- 18« —0— (9CP- 270» |
0.5 1

A
0.4
Ghanas of
Change of /A \/ Inside
InsidB °-3
Diameter
m
K Diamter
00 ,

/-^ //
0.2

0.1 rf r^ M
/

0.5

0 ^L V
3 20 40 60 80 100
20 40 60
Location (X, 0:Top — IQCKBotto»)
80

Location Of, 0:Top — 100:Botton)

Figure 12. Change of ID distribution of used IN519 Figure 13. Relationship between ID change and
catalyst tube for 19 years in plant B. hardness distribution in used IN519 catalyst tubes
for 20 years in plant A.

However, if the initial dimensions before tube use and microstructures and aging conditions arranged accord-
the location of the maximum tube skin temperature are ing to an appropriate aging parameter (such as the
known, it is not necessary to measure all parts of the cat- Larson-Miller parameter) is available, it is possible to
alyst tube. It is enough to check the dimensions on the estimate the service temperature of the catalyst tube
maximum tube-skin temperature portion. from the microstructure observation.
A method of measuring the OD and tube wall thick- The creep damage of used catalyst tube material can
ness for used catalyst tubes at the plant site was pro- also be confirmed through microscopic observation of
posed and automatically performed with ultrasonic test- the distribution of creep voids and fissures formed at
ing and other equipment (Shannon and Smith, 1997). the interface between matrix and primary carbides
and/or sigma phases of grain boundaries.
Observation of Microstructure
Guide for Remaining Life Assessment
The microstructure of catalyst tube changes with
exposed temperature and time the tube has been in Figures 13 and 14 show the relationship between ID
use. If the database that shows the relation of the change and hardness distribution of used IN519 cata-

AMMONIA TECHNICAL MANUAL 160 1999

 —o— Change of innac diameter (O*- 18O [ [ —O— Haritiass (Center) [
2

O O
Al F
1.5
A2
Changs of
Inside
Diameter E
00 »
U
L2 O
0.5 O
O
U •X °c

20 40 100 0
1M 140 160 180 200 220
Location Of, 0:Top — 100:Botton>
Hardness (Hv)

Figure 14. Relationship between change of ID and Figure 15. Relationship between hardness and
hardness distribution of used IN519 catalyst tube change of ID of used IN519 catalyst tubes in plants
for 19 years in plant B. AtoM.

lyst tubes in plant A and plant B. The change of ID change of ID and the estimated creep rupture time at
increases as it goes downstream. The portion of the 1173 K and 13.7 MPa for the test coupon sampled
tube that exhibits the maximum diameter increase cor- from the maximum service temperature portion in
responds to the portion of the lowest hardness. This is plants A through L. The conditions of temperature and
also the portion of the tube that has experienced the stress for the estimation were selected for the typical
highest service temperature and the most severe creep service condition of those catalyst tubes. The estimat-
damage. Figure 15 shows the relationship between the ed remaining life of the tube decreases as ID increases.
hardness and the change of ID at the maximum service The catalyst tube used for 19 years hi plant B (noted B
temperature portion of used IN519 catalyst tubes in hi Figure 17), however, had remaining life of 30,000
service for more than 12 years in plants A to M. h, although its change of ID was less than 0.5%. On
Figure 16 shows the result of relationship between the other hand, the catalyst tubes used for 20 years in
the change of ID and the tube-life consumption at the plant A (noted Al and A2 hi Figure 17) had remaining
maximum service temperature portion for the IN519 life of 15,000 and 21,000 h with ID changes of 1.7 and
catalyst tube in plant A. The life of the EST519 catalyst 1.5%, respectively. The difference of the change of ID
tube hi plant A was calculated by adding actual service between plants B and plant A is inferred on the basis
time (20 years) to the estimated rupture tune under the of the service temperature difference. The service tem-
service condition, namely remaining life of the tube, perature of the catalyst tube hi plant A is higher than
based on the results of creep rupture tests, which were that of the tube hi plant B. The results of hardness dis-
performed on the test coupons sampled from the used tribution measurements of the tubes used hi plant A
IN519 catalyst tube material for 20 years hi plant A. and B are shown hi Figures 3 and 4.
The remaining life of the tube was about 2 years. The Therefore, when the remaining life of catalyst tube
life consumption was calculated from the actual ser- is evaluated by dimensional change, the effect of ser-
vice tune divided by the life of the tube, which is 22 vice temperature on creep deformation should be con-
years (20 + 2 years). Those tubes had been used for sidered. The quantitative relationship between the
20 years, and then were replaced with HP-Nb,Ti cata- creep deformation and the remaining life of the tube
lyst tubes before rupture. The 1.7% bulging of ID is material is estimated from the creep data and the creep
considered as 95% life consumption (20 years/ 22 rupture data of used catalyst-tube material. To com-
years). plete the quantitative information for creep deforma-
Figure 17 shows the relationship between the tion, many creep test data will be analyzed and the

AMMONIA TECHNICAL MANUAL 161 1999

 3.5 10s

Ur
ZS Estimated JD_
Change Creep
of ID 2 Rupture s °C
00 Tine at 10
1173 K,
1.5 13.7 «Pa

10"
0.5 1 1.5
0.4 0.6 0.8 Change of Inside Diameter 00
Life consumption

Figure 16. Relationship between change of ID and Figure 17. Relationship between change of ID and
Me consumption of used IN519 catalyst tube for 20 remaining life of used IN519 catalyst tubes in plants
years in plant A. AtoL.

relationship between creep strain, time, temperature, confirm the metal temperature and the creep damage.
and stress will be obtained. If the application of • A 1.7% change of inside diameter of IN519 cata-
remaining tube-life assessment were limited to IN519 lyst tube at the highest service-temperature portion as
catalyst tubes, the 1.7% change of ID at the highest a possible deciding factor in tube replacement.
service temperature portion might be the guideline to The same procedure can be applied to other catalyst
indicate the need for replacement within 2 years. tube material, such as microalloyed HP-Nb material.

Conclusion Literature Cited

We propose the following procedure for assessing Shannon, B. E., and N. Smith, "Assessing Creep
remaining tube life of catalyst tubes by dimensional Damage in Cast Furnace Tubes Using
check. Nondestructive Examination Techniques," Ammonia
• A measurement of hardness distribution to find the Plant Safety & Related Facilities, Vol. 35, AIChE,
temperature profile of catalyst tube. New York (1998).
• A measurement of OD and tube wall thickness at Shibasaki, T., T. Mohri, and K. Takemura, "Remaining
the same position to evaluate the amount of creep Life Assessment of Nb-Containing Catalyst Tubes
damage. for Steam Reformer Furnace," Ammonia Plant
• An observation of microstructures from the outer Safety & Related Facilities, Vol. 33, AIChE, New
surface side of the catalyst tube by replica method to York (1996).

AMMONIA TECHNICAL MANUAL 162 1999

 DISCUSSION
Brian Shannon, lesco Inc. : In any of the items that naces you were studying, in other words, where the
you have done your analysis on, have you seen any of burners are located?
the samples that had creep or dimensional damage that Mohri: I'm responsible for design of steam reformers
did not have internal microstructure deformation? If for Chiyoda Corp. Your question concerns the type of
the tube had 1.7% of creep by measurement, were reformer furnace. This type is the Haldor Tops0e radi-
there any occasions where it was not damaged inter- ant wall type hydrogen reformer, and we, of course,
nally? analyze the firing pattern for the reformer tube itself,
Shibasaki: Yes. By microscopic examination, we and we find out some remarkable things. This
found creep voids and fissures in the matrix which reformer has six rows of side wall burners. Especially,
was deformed more than 1% to 1.5%. We tried to the lower three burners were firing more than the
detect by UT measures, but we could not find a design, so this means the firing pattern did not follow
change of ultrasonic properties with this method. exactly the design heat input. I'm not sure of the rea-
Robert R. Jungerhans, Kinetics Technology son why the client has been heating more than the bot-
International Corp.: It appears from your graphs that tom portion of the catalyst tube. This is a remarkable
the maximum skin temperature is always towards the fact of this reformer. However, in other side wall fired
bottom, towards the outlet part of the tube. My ques- reformers in Japan designed by Tops0e and Chiyoda,
tion is related to the heat release pattern in the fur- the heat input was more even on tube length.

AMMONIA TECHNICAL MANUAL 163 1999